chore: 并入 iotlcm-bot 历史到 bot/（原提交 31dbc1d）
This commit is contained in:
MiFaZhan
@@ -0,0 +1,198 @@
|
||||
package notify
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
"log/slog"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"git.5iot.com/iotlcm-public/iotlcm-bot/internal/state"
|
||||
)
|
||||
|
||||
func writeArchive(t *testing.T, dir, name, body string) {
|
||||
t.Helper()
|
||||
if err := os.WriteFile(filepath.Join(dir, name), []byte(body), 0o600); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
|
||||
func discardLogger() *slog.Logger {
|
||||
return slog.New(slog.NewTextHandler(os.Stderr, &slog.HandlerOptions{Level: slog.LevelError + 1}))
|
||||
}
|
||||
|
||||
// countingLogger 统计 ERROR 级日志条数，用于验证去重告警。
|
||||
func countingLogger(count *int) *slog.Logger {
|
||||
return slog.New(slog.NewTextHandler(countingWriter{count}, &slog.HandlerOptions{Level: slog.LevelError}))
|
||||
}
|
||||
|
||||
type countingWriter struct{ count *int }
|
||||
|
||||
func (w countingWriter) Write(p []byte) (int, error) {
|
||||
*w.count++
|
||||
return len(p), nil
|
||||
}
|
||||
|
||||
const validArchiveBody = `{"version":1,"run":{"runId":"z-good","operation":"op","status":"completed","successCount":1,"failedCount":0,"createdAt":"2026-10-12T00:00:00Z","completedAt":"2026-10-12T00:01:00Z"},"rounds":[{"round":1,"enqueuedCount":1,"activeCount":0,"nextRoundCount":0,"successCount":1,"failedCount":0,"startedAt":"2026-10-12T00:00:00Z","completedAt":"2026-10-12T00:01:00Z"}]}`
|
||||
|
||||
// 一个坏归档不能挡住后面的归档，但也不能每轮扫描都刷一条日志。
|
||||
func TestScanSkipsBrokenArchiveAndLogsOnlyOnce(t *testing.T) {
|
||||
root := t.TempDir()
|
||||
// 字典序在前：旧格式（rounds 为空）会校验失败。
|
||||
writeArchive(t, root, "run-broken.json", `{"version":1,"run":{"runId":"broken","operation":"op","status":"completed"},"rounds":[]}`)
|
||||
writeArchive(t, root, "run-z-good.json", validArchiveBody)
|
||||
|
||||
fake := &fakeFeishu{}
|
||||
var logged int
|
||||
svc := Service{archiveDir: root, environment: "preview", store: state.New(t.TempDir()), client: fake, logger: countingLogger(&logged), warned: map[string]struct{}{}}
|
||||
|
||||
for i := 0; i < 3; i++ {
|
||||
if err := svc.Scan(context.Background()); err != nil {
|
||||
t.Fatalf("Scan() returned error for a broken archive: %v", err)
|
||||
}
|
||||
}
|
||||
if fake.sent != 1 {
|
||||
t.Fatalf("expected the valid archive to be sent once, sent=%d", fake.sent)
|
||||
}
|
||||
if logged != 1 {
|
||||
t.Fatalf("broken archive should be reported once across 3 scans, logged=%d", logged)
|
||||
}
|
||||
}
|
||||
|
||||
// 文件被重写（内容变了）后应该重新告警。
|
||||
func TestScanWarnsAgainAfterArchiveIsRewritten(t *testing.T) {
|
||||
root := t.TempDir()
|
||||
writeArchive(t, root, "run-broken.json", `{"version":1,"run":{"runId":"broken","operation":"op","status":"completed"},"rounds":[]}`)
|
||||
|
||||
var logged int
|
||||
svc := Service{archiveDir: root, environment: "preview", store: state.New(t.TempDir()), client: &fakeFeishu{}, logger: countingLogger(&logged), warned: map[string]struct{}{}}
|
||||
|
||||
if err := svc.Scan(context.Background()); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
// 重写为另一个仍然无效的内容，mtime/大小变化
|
||||
time.Sleep(10 * time.Millisecond)
|
||||
writeArchive(t, root, "run-broken.json", `{"version":1,"run":{"runId":"broken","operation":"op","status":"completed","queryMonth":"202610"},"rounds":[]}`)
|
||||
if err := svc.Scan(context.Background()); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if logged != 2 {
|
||||
t.Fatalf("rewritten broken archive should be reported again, logged=%d", logged)
|
||||
}
|
||||
}
|
||||
|
||||
// 发送失败（飞书临时故障）必须每轮都报，不能被去重掉：
|
||||
// 否则一次网络抖动就会让通知静默丢失。
|
||||
func TestScanSendFailureIsNotDeduped(t *testing.T) {
|
||||
root := t.TempDir()
|
||||
writeArchive(t, root, "run-z-good.json", validArchiveBody)
|
||||
|
||||
fail := &fakeFeishu{err: errors.New("feishu timeout")}
|
||||
var logged int
|
||||
svc := Service{archiveDir: root, environment: "preview", store: state.New(t.TempDir()), client: fail, logger: countingLogger(&logged), warned: map[string]struct{}{}}
|
||||
|
||||
for i := 0; i < 3; i++ {
|
||||
if err := svc.Scan(context.Background()); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
}
|
||||
if logged != 3 {
|
||||
t.Fatalf("transient send failures must be reported every scan, logged=%d want=3", logged)
|
||||
}
|
||||
}
|
||||
|
||||
func TestHealthReportsEmptyDirectory(t *testing.T) {
|
||||
svc := Service{archiveDir: t.TempDir()}
|
||||
h := svc.Health()
|
||||
if !h.Mounted || h.Degraded {
|
||||
t.Fatalf("empty but mounted directory should not be degraded: %+v", h)
|
||||
}
|
||||
if h.Entries != 0 || h.Reason == "" {
|
||||
t.Fatalf("unexpected health: %+v", h)
|
||||
}
|
||||
}
|
||||
|
||||
func TestHealthReportsMissingDirectory(t *testing.T) {
|
||||
svc := Service{archiveDir: filepath.Join(t.TempDir(), "gone")}
|
||||
h := svc.Health()
|
||||
if h.Mounted || !h.Degraded {
|
||||
t.Fatalf("missing directory must be degraded: %+v", h)
|
||||
}
|
||||
}
|
||||
|
||||
// 挂载失效时必须退出进程，交给 restart policy 重新拉起以重新解析 bind mount。
|
||||
// 这是「不再需要人工同批重建 bot」的关键：重启后挂载指向宿主当前目录。
|
||||
func TestWatchExitsWhenArchiveMountIsDegraded(t *testing.T) {
|
||||
shortenExitDelay(t)
|
||||
if !ExitOnDegraded {
|
||||
t.Skip("ExitOnDegraded disabled")
|
||||
}
|
||||
stateDir := t.TempDir()
|
||||
svc := Service{
|
||||
archiveDir: filepath.Join(t.TempDir(), "gone"),
|
||||
environment: "test",
|
||||
stateDir: stateDir,
|
||||
pollInterval: 10 * time.Millisecond,
|
||||
store: state.New(stateDir),
|
||||
logger: discardLogger(),
|
||||
}
|
||||
done := make(chan error, 1)
|
||||
go func() { done <- svc.Watch(context.Background()) }()
|
||||
select {
|
||||
case err := <-done:
|
||||
if err == nil {
|
||||
t.Fatal("Watch must return an error so the process exits non-zero and restarts")
|
||||
}
|
||||
case <-time.After(3 * time.Second):
|
||||
t.Fatal("Watch did not exit after detecting a degraded archive mount")
|
||||
}
|
||||
}
|
||||
|
||||
// 目录长期为空时才缩短轮询；有内容或空窗很短时保持配置周期。
|
||||
func TestIntervalShortensOnlyAfterLongEmptyPeriod(t *testing.T) {
|
||||
root := t.TempDir()
|
||||
svc := Service{archiveDir: root, pollInterval: time.Minute}
|
||||
if got := svc.Interval(time.Time{}); got != time.Minute {
|
||||
t.Fatalf("fresh empty dir should keep poll interval, got %s", got)
|
||||
}
|
||||
if got := svc.Interval(time.Now().Add(-StaleAfter - time.Second)); got != 10*time.Second {
|
||||
t.Fatalf("long-empty dir should poll every 10s, got %s", got)
|
||||
}
|
||||
writeArchive(t, root, "run-anything.json", "{}")
|
||||
if got := svc.Interval(time.Now().Add(-time.Hour)); got != time.Minute {
|
||||
t.Fatalf("non-empty dir should keep poll interval, got %s", got)
|
||||
}
|
||||
}
|
||||
|
||||
type fakeFeishu struct {
|
||||
sent int
|
||||
err error
|
||||
textCount int
|
||||
texts []string
|
||||
}
|
||||
|
||||
func (f *fakeFeishu) Send(ctx context.Context, card []byte) error {
|
||||
if f.err != nil {
|
||||
return f.err
|
||||
}
|
||||
f.sent++
|
||||
return nil
|
||||
}
|
||||
func (f *fakeFeishu) SendText(ctx context.Context, text string) error {
|
||||
f.textCount++
|
||||
f.texts = append(f.texts, text)
|
||||
return nil
|
||||
}
|
||||
|
||||
// newTestStore 构造一个 state.Store 并返回其 Dir，供 alertstate 测试使用。
|
||||
func newTestStore(dir string) state.Store { return state.New(dir) }
|
||||
|
||||
// shortenExitDelay 把退出等待改短，避免测试白等 30 秒（仅测试内使用）。
|
||||
func shortenExitDelay(t *testing.T) {
|
||||
t.Helper()
|
||||
original := DegradedExitDelay
|
||||
DegradedExitDelay = 10 * time.Millisecond
|
||||
t.Cleanup(func() { DegradedExitDelay = original })
|
||||
}
|
||||
Reference in New Issue
Block a user